376 



Fishery Bulletin 105(3) 



tic. Locales farther north in the Atlantic lack the Gulf 

 haplotypes; in this scenario it would be expected that 

 Atlantic menhaden collected farther north would lack 

 the Gulf menhaden influence. Indeed, in the divergent 

 group of Atlantic menhaden examined in the present 

 study, four of the eight haplotypes came from juve- 

 niles collected from the northernmost sampling locale 

 (Maine). The remaining haplotypes came from a single 

 adult from a northern locale (New Jersey), two adults 

 from southern locales (North Carolina), and a southern 

 locale juvenile (North Carolina). 



Menhaden 



0.00 



0.05 



0.08 



B Small-scaled menhaden 



0.00 



0.05 



0.10 



0.15 



0.10 

 0.08 

 0.06 

 0.04 

 0.02 

 0.00 



D Clade I large-scaled menhaden 



0.10 



0.20 



0.00 0.02 0,04 0.06 0.08 



Genetic distance 



Figure 5 



Saturation plots of menhaden mtDNA data. The x-axis represents 

 Kimura (1980) distances between sequences, whereas the v-axis rep- 

 resents the rate of transitions (♦) and transversions (■). A saturation 

 plot was constructed for (A) mtDNA samples from all four species, 

 (B) samples from small-scaled menhaden only (yellowfin menhaden 

 [Brevoortia smithi] and finescale menhaden [B. gunteri]), (C) samples 

 from large-scaled menhaden only (Gulf menhaden \B. patronus] and 

 Atlantic menhaden [B. tyrannus]}, and (D) large-scaled menhaden 

 from mtDNA clade I only. 



Previous treatments have recognized two subpopula- 

 tions ofS. smithi in the Gulf and Atlantic, and Hildeb- 

 rand (1963) characterized the eastern Gulf B. smithi 

 as intermediate in form between B. gunteri and the 

 Atlantic form of B. smithi. An interesting note in our 

 data set is the large divergence estimate between two 

 sampled populations of JS. sm.ithi obtained in the micro- 

 satellite analysis. This estimate of divergence between 

 these populations (F,,'''"'"" = 0.0523) is the largest of 

 any comparison and is an order of magnitude greater 

 than the mean comparison among populations of the 

 same species iF^f'""'" = 0.0054). Brevoortia 

 smithi is the only species of menhaden 

 that has significantly large populations in 

 both the Gulf and Atlantic, and popula- 

 tions of this species are rare towards the 

 southern tip of Florida (Reintjes, 1959; 

 Dahlberg, 1970). Thus, vicariance may 

 play a role in divergence between these 

 populations. It is clear that biological dif- 

 ferences between Atlantic and Gulf forms 

 of B. smithi extend into the genetic data 

 presented here, resulting in long branches 

 in the microsatellite-based topology. In ad- 

 dition, the smaller relative census sizes of 

 small-scaled menhaden species has likely 

 resulted in a faster and more distinctive 

 pattern of divergence among populations 

 of these species than what has occurred in 

 large-scaled species. Further investigation 

 is needed in order to determine if the ge- 

 netic divergence between Gulf and Atlantic 

 forms of B. smithi is genome-wide. 



A final peculiarity of the data was a 

 single individual B. gunteri which had a 

 genotype that indicated influence from 

 B. patronus (Fig. 3). Specifically, this 

 individual had an estimated 72.3% of its 

 genotype contributed by a cluster roughly 

 representative of B. patronus, whereas 

 all other individuals of this species had 

 less than a 10% contribution from this 

 cluster (averaged around 2.7%). Although 

 high rates of hybridization between B. 

 smithi and both large-scaled menhaden 

 species have been documented in pen- 

 insular Florida (Turner, 1969; Hettler, 

 1984), hybrids between B. gunteri and 

 B. patronus have not been documented 

 (Anderson and McDonald, 2007). Visual 

 inspection of the individual in question 

 yielded no morphological evidence of hy- 

 bridization. In addition, Bayesian assign- 

 ment can be expected to perform better 

 if more loci are sampled. Although five 

 microsatellite loci were adequate to re- 

 cover evidence for population structure 

 in the present data set, individual as- 

 signments based on limited genetic loci 

 will comprise a high level of uncertainty 



